西南交通大学学报

Short-Term Load Forecasting Based on Complex Morlet Wavelet SVM

CHEN Weirong, ZHENG Yongkang, DAI Chaohua, WANG Weibo

2009, 22(5): 631-636.

[Abstract](1771)

Abstract:
In view of the advantage of wavelet analyses in subtle feature extraction and the good global optimization ability of cloud theory-based genetic algorithm (CGA),a CGA-based complex Morlet wavelet SVM (support vector machine),called CGA-CMW-SVM for short,was proposed to improve the forecasting precision and easily select the parameters of SVM.In the CGA-CMW-SVM,the complex Morlet wavelet is used as the kernel function,and the CGA is adopted to optimize the parameters.To decrease the system complexity in short-term load forecasting,the load time series were reconstructed based on the phase space reconstruction theory and their chaotic characteristics only by considering the historical load data without other factors,such as weather and holidays.Though it is believed that the single load data is often characterized as incomplete and inaccurate information,the phase space reconstruction can overcome the shortcomings.Then,the phase space vectors were used as the inputs of the CGA-CMW-SVM for short-term load forecasting.The simulation experiments show that the presented method has small average and maximum errors,and its average errors are less than 1.3400% with a minimum value of 1.0087%.

Incremental Clustering Algorithm Based on Rough Reduction for Data Stream

YIN Guisheng, YU Xiang, NING Hui

2009, 22(5): 637-642,653.

[Abstract](1508)

Abstract:
An incremental algorithm based on rough reduction for clustering data stream,named as RICStream (rough incremental clustering stream),was proposed to overcome the shortcoming of the CluStream algorithm,i.e.,the number of clusters must be predefined when it is used and the evolution of data stream can not be reflected efficiently to influence clustering results.With the guarantee of clustering accuracy,the RICStream adjusts the attributes of data stream incrementally to lead to the reduction of clustering time and computation cost.In order to store data stream a novel grid structure which can be adjusted incrementally was put forward.As a result,the changes of data stream can be efficiently reflected by clustering results.The experiments on real datasets and synthetic datasets show the applicability and validity of the RICStream.

Improved Decision Tree Algorithm Based on Samples Selection

FENG Shaorong, XIAO Wenjun

2009, 22(5): 643-647.

[Abstract](1519)

Abstract:
To raise the accuracy of decision tree classification algorithms,an improved decision tree classification algorithm based on samples selection was proposed by comparing several classical decision tree classification algorithms.This improved algorithm searches better samples through a constantly iterative process based on the facts that the correlation between decision trees’ accuracy and samples is large and decision trees can only get a local optimal solution.As a result,a better decision tree classification algorithm can be obtained under the condition of not changing the decision tree classification algorithm.The improved algorithm is not aiming at a decision tree and it carries through iteration only based on some feedback information of input and output,so its universality is better.Experimental results show that the ratio of the average error rates of the improved algorithm and the ID3,C4.5 algorithms is about 0.82 to 1.22 to 0.92.

Planar IGBT Fabrication and Forward Voltage Drop Calculated by PIN Model

YUAN Shoucai, LIU Yamei

2009, 22(5): 648-653.

[Abstract](2114)

Abstract:
IGBT (insulate gate bipolar transistor) devices with planar processes were designed and fabricated based on planar silicon CMOS (complimentary metal oxide semiconductor)technology equipments and epitaxial silicon substrates.The measured output current-voltage characteristics of the fabricated IGBT devices are excellent,and their threshold voltage is from 3.7 to 4.8 V.The forward voltage drops of the IGBT devices were analyzed and calculated by PIN diode (a p-type near intrinsic and an n-type semiconductor) model with their designed structure and process parameters,and the calculated forward voltage drops were compared with the measured data.The comparison result shows that they are fitting better when conducted currents are small,but there exist differences between the calculated and measured forward voltage drops due to model simplification and parasitic series resistors in the test systems when conducted currents are large and the calculated forward voltage drops are about 8% smaller than the measured values under the same drain-source conducted currents.

Design of Global Sliding-Mode Controlled Buck Converters Based on Equivalent Control

NI Yu, XU Jianping

2009, 22(5): 654-659.

[Abstract](1621)

Abstract:
To simplify the design of sliding-mode controller and improve the robustness of converter systems,the principle and design method of global sliding-mode control were investigated and introduced to the controller design for Buck converters.The main characteristics of Buck converters with global sliding-mode control were analyzed,the design procedure of global sliding-mode controller was given based on the equivalent control idea,and global sliding-mode control law was designed.Finally,a simulation was carried out to explore the feasibility and validity of applying the global sliding-mode controller to Buck converters.The simulation results show that a Buck converter with the global sliding-mode controller has a high transient response velocity than a Buck converter with the conventional sliding-mode controller,and the former possesses global robustness throughout the transient process.

Control Laws and Characters of Pulse Train Controlled Buck Converters

QIN Ming, XU Jianping, MU Qingbo

2009, 22(5): 660-666.

[Abstract](1502)

Abstract:
The merits and implementation of voltage mode pulse train (VMPT) and current mode pulse train (CMPT) control techniques for switching DC-DC converters were analyzed by theoretical analysis,simulation and experiment to compare their steady state characters,control regulation and startup process.The results show that the control regulations of VMPT and CMPT controllers have basically the same control law under the variation in load,i.e.they generate more high-energy pulses when the load increases;under the variation in input,i.e.when the input voltage increases,VMPT and CMPT controllers generate more low-energy and more high-power pulses,respectively;and the startup process of the CMPT converter is smoother compared with that of the VMPT converter.

Integrity Check Method for Fine-Grained Data Based on Complex Rotary Codes

CHEN Long, FANG Xinlei, WANG Guoyin

2009, 22(5): 667-671.

[Abstract](1552)

Abstract:
To overcome the shortcoming of the existing integrity check methods for fine-grained data and improve the error indication capability,a novel multiple-error indicating method based on the complex rotary codes was proposed.This method checks the integrity of data objects via cross-checking with grouped Hashes,where each group Hashes can indicate the integrity of all data independently.Analysis and experiment results show that the proposed method can compress Hash data effectively,and its base error amplification ratio reaches almost 1,much lower than that of the existing check methods.

Fragile Watermarking Algorithm Based on Circular Structure

SONG Wei, HOU Jianjun, LI Zhaohong

2009, 22(5): 672-676.

[Abstract](1486)

Abstract:
To improve the ability of tamper localization of watermarking,a novel fragile watermarking algorithm based on circular structure was proposed.In this algorithm,images are divided into blocks,and the blocks are modulated by the Logistic mapping to generate block watermarks.Then cat mapping is utilized to construct a circular structure in order to build a one-to-one circular relationship between image blocks.The block watermarks are embedded into LSB (least significant bit) of the corresponding blocks in light of the circular structure.The simulation results show that the proposed algorithm can not only localize the tampers correctly and accurately,but also resist the statistic attack and the counterfeiting attack.

Reconstruction of Implicit Surfaces from Uncalibrated Camera Image Sequences

JIANG Yongquan, PENG Qiang, CHEN Jinxiong

2009, 22(5): 677-681,687.

[Abstract](1614)

Abstract:
To improve the quality of 3D geometric models reconstructed from uncalibrated image sequences,a feature tracking algorithm was proposed to match more points among a sequence of images.In the algorithm,a dense grid in the first frame is drawn,and easily tracked feature points near the grid points are determined.The sub-pixel coordinates of the feature points are found by iteration.The optical flow for this sparse feature set is calculated using iterative Lucas-Kanade method in pyramids.An even and dense 3D point-cloud is reconstructed with a self-calibration algorithm.Finally,an RBF (radial basis function) implicit surface reconstruction was applied to generate the surface model of the object.By experimenting with a number of image sequences,the reconstruction results show that the algorithm can obtain satisfactory surfaces for images with rich texture.

LMI-Based Robust Stabilization of Linear Uncertain Systems with Input Time-Delay

ZHANG Yougang, LUO Zhiyong

2009, 22(5): 682-687.

[Abstract](1400)

Abstract:
To robust stabilization problems for a linear uncertain system with parametric uncertainties,state delay and input delay were considered,where the parametric uncertainties are norm-bounded,and the state and input delays are time-varying and bounded.For robust stabilization of this system,a memoryless state-feedback robust stabilization controller was designed,and LMI-formed (linear matrix inequalities) sufficient conditions for robust exponential stability of the closed-loop system under the controller were given.These sufficient conditions don not require that the derivative of delays be less than 1,so they are less conservative than those in the current researches.Furthermore,they can be applied in both fast-and slow-varying time delay systems.Finally,a numerical example was given to demonstrate the validity of the proposed robust stability conditions.

Stability and L2-Gain of Linear Parameter-Varying Subsection System

WANG Tao, LI Jikun, LIU Li

2009, 22(5): 688-692.

[Abstract](1692)

Abstract:
In order to reduce conservative performances resulted in by the application of the Lyapunov function to the analysis of a linear parameter-varying system,the relationships of the stability and L2-gain of the system with the subsection and changing rate of parameters were investigated.By considering the subsection and changing rate of parameters together,a design method for linear parameter-varying systems was put forward based on LMI (linear matrix inequality).Simulation results show that the system is robust and external disturbance is outstandingly restrained to prove this method applicable to all kinds of linear parameter-varying systems.

Linear FM Signal Estimation Method Based on Linear Iterative Prediction

XU Zhengwu, LU Xiaoqian, ZHANG Yang

2009, 22(5): 693-696.

[Abstract](1464)

Abstract:
To enlarge the parameter estimation range in the algorithm for estimation of LFM (linear frequency modulation) signal parameters,and improved algorithm was proposed based on the method of rotate-add-decimate (RAD) for frequency measurement,and it was implemented with the method of linear iterative prediction.It increases the estimation range to (-π/2,π/2) on the premise of ensuring linear FM (frequency modulation) ratio and estimation precision.Simulation demonstrated the effectiveness of the algorithm.

Damage Mechanism of Reinforced Concrete Frames in Fire

FU Chuanguo, WANG Guangyong, WANG Yuzhuo

2009, 22(5): 697-703,709.

[Abstract](1695)

Abstract:
Based on software ABAQUS,a user material model and subroutine for the analysis of reinforced concrete frame structures in a fire were developed to simulate the cross-section temperature field of reinforced concrete components more accurately.With the incremental theory of plasticity,the non-linear analysis of a reinforced concrete frame model subjected to a dead load and a local fire at different conditions were carried out.Based on the non-linear analysis results,the distribution of plastic hinges and the damage mechanism of the reinforced concrete frame in a fire were revealed,and the effects of fire room location on the failure mechanism of reinforced concrete frames were discussed.In the end,the most dangerous location of fire room was explored.The analysis results indicate that the reinforced concrete frame members undergo internal force redistribution under a fire,and the superposition of constraint internal forces with external loads causes plastic hinges in beams and columns.In addition,the distribution of plastic hinges and the damage mechanism of reinforced concrete frames are affected obviously by fire room location,so the disadvantage of fire room location should be considered in a fire resistant design.

Parameter Identification of Long-Span Cable-Stayed Bridge Based on Grey-Neural Network

MENG Qingcheng, QI Xin, LI Qiao, BU Yizhi

2009, 22(5): 704-709.

[Abstract](1807)

Abstract:
To solve the structural parameter identification problem with measurement noises during the construction process of long-span cable-stayed bridges,an entire-processed parameter identification method based on grey-neural network was proposed.The fusion of the grey system theory and the artificial neural network makes it possible to identify and forecast structural parameters even under the condition of insufficient samples or incomplete data.The identification results for the Su-Tong bridge show that with the proposed method the maximum error between post-identification result and real structural response decreases by 77% and the estimation accuracy is raised by 50% than with traditional ones.

Torsion Analysis of Horizontal Curved Box Girder under Temperature Difference

FU Chunyu, LI Qiao

2009, 22(5): 710-715.

[Abstract](1643)

Abstract:
In order to calculate the internal forces and displacements of horizontal curved box girders under temperature difference correctly,the torsion deformations caused by section rotation and temperature difference were analyzed based on the characteristic of the flexure-torsion coupling,and the torsion and deflection of simply-supported horizontal curved box girders were calculated.By taking a simply-supported horizontal curved box girder as an example of simple structures,a method to calculate internal forces in continuous box girders was put forward.The internal forces in a continuous curved box girder with two equal spans were calculated with this method and compared with the results of 3-D finite element analysis to show an error of less than 5%.The research shows that when the displacement and internal force of horizontal curved box girders are calculated,the torsion deformation caused by temperature difference should be taken in account.

Influences of Lining Stiffness Distribution on Frost Heave Pressure on Cold Regional Tunnels

DENG Gang, WANG Jianyu, ZHENG Jinlong

2009, 22(5): 716-720.

[Abstract](1706)

Abstract:
To ensure that lining has a suitable safety degree,a constrained frost heave model was proposed based on the interaction characteristic of frozen water body,lining and surrounding rock,and the corresponding analytical expression of frost heave pressure was give.A load-structure model was adopted to analyze the distribution of lining stiffness.In addition,the relationship between frost heave pressure and lining stiffness was investigated by taking the structure safety degree as an assessment index.The research result shows that the stiffness distribution of lining with identical thickness is uneven-lining stiffness is small at the arch crown of a tunnel and middle part of its inverted arch,while it is large along its side walls.The increase in the thickness of lining will make lining stiffness raise to result in a great frost heave pressure and the load-bearing capacity of lining increase significantly.As a result,increasing the thickness of lining is still an effective measure to improve the anti-freeze ability of cold regional tunnels.

Fractal Description of Rail Corrugation and Its Dynamic Simulation

WU Zeng, LIU Xueyi, YAO Lingkan

2009, 22(5): 721-725.

[Abstract](1653)

Abstract:
To probe into the method to accurately describe the attrition grade of rail corrugation using fractal dimension,the characteristics of rail corrugation,measured in some segments with a curve radius of 300 m along the Shijiazhuang-Taiyuan railway line,were described by fractal dimension based on experimental data,and its fractal dimension was calculated out.The trace curve of rail corrugation was inputted as track cross-level irregularity,and the vertical displacement,acceleration and other dynamic responses of components under different train speeds were calculated numerically,and the relationship between the dynamic responses and the fractal dimension was investigated.The research shows that the attrition grade of rail corrugation can be described by its fractal dimension and the dynamic responses calculated out and the fractal dimension are influenced by measuring distance

Nonlinear Beam Element with Meshed Sections

ZHOU Lingyuan, LI Tongmei, LI Qiao

2009, 22(5): 726-731.

[Abstract](1722)

Abstract:
Based on the distributed plasticity approach,a new kind of beam element,3-D materially nonlinear beam element with meshed sections,was presented to analyze the materially nonlinear behaviour of frame structures accurately,and a numerical example for the elasto-plastic analysis of a cantilever beam was given to illustrate the accuracy and efficiency of this approach.With this approach,beam sections are discretized into planar isoparametric elements at the integral points along the longitudinal direction of beam element,the stiffness matrices and resistant forces of sections are calculated by introducing the nonlinear stress-strain relations of materials and respectively integrating the stress-strain relations and stresses of sections,and the element stiffness and nodal resistant forces can also be evaluated with displacement interpolation functions when materials are in a nonlinear state.

Finite Element Model for Catenary in Initial Equilibrium State

LI Ruiping, ZHOU Ning, MEI Guiming, ZHANG Weihua

2009, 22(5): 732-737.

[Abstract](1648)

Abstract:
Finite element models were established using link and beam elements,respectively,to accurately and quickly calculate the dropper lengths,static stiffness,natural frequencies and mode shapes of catenaries in the initial equilibrium state.The results show that the two element types exert no obvious effect on the static and dynamic simulation results;the calculation efficiency can be improved using the link elements due to their fewer freedom degrees;the equilibrium positions and the axial stiffness of the elements should be determined in advance when the static stiffness and mode shapes are calculated.

Coupled Finite Element Thermal-Mechanical Analysis of Gravity Support System of International Thermonuclear Experimental Reactor

LIANG Shangming, YAN Xijiang, WANG Xianzhou

2009, 22(5): 738-742.

[Abstract](2079)

Abstract:
A method for building the finite element model of the gravity support system of the international thermonuclear experimental reactor (ITER) was put forward according to the structure characteristics of the gravity support system.By this method,a 3D finite element model of the gravity support system was built with ANSYS software using a mesh generation method of high precision and with an acceptable calculation scale.Then,a coupled thermal-mechanical analysis of the gravity support system was carried out,and the maximum Tresca stress and its position in every part of the system were obtained under the combined action of dead weight,thermal load,electromagnetic forces and seismic load.The stress analysis results show that the maximum stresses in all parts of the system are smaller than their respective allowable stress limits,and that the parts have enough mechanical strength to meet the requirements of ASME Pressure Vessel Code-III NF.

Experimental Investigation on Periodical Temperature Field in Active Magnetic Regenerator of Rotary Magnetic Refrigerator

WU Jianghong, WANG Xihui, TANG Yongbo, CHEN Yungui

2009, 22(5): 743-747,793.

[Abstract](1496)

Abstract:
The heat transfer rules between the porous medium and single-phase fluid under the periodical instantaneous heat source in the active magnetic regenerator (AMR) of a rotary magnetic refrigerator were discussed based on temperature field measurement.The experimental results showed that the temperature vibration frequency in the AMR was equal to the rotational frequency of the AMR.With the increasing of the rotational frequency of the AMR,the temperature vibration frequency increased accordingly,but the temperature values decreased.On the other hand,the temperature vibration frequencies in demagnetizing area and magnetizing area were the same,the vibration phase shift between the two areas was about 180 degrees,and the vibration amplitude of the former area was nearly twice as lange as that of the latter one.Generally,increasing the rotational frequency of the AMR has a positive effect on the refrigeration capacity,but the temperature reached its climax as the fluid flow increased to 2.639×10-4 m3/s,so the heat transfer fluid flow should designed to be less than the critical value.

Adaptive Semiactive Bumper for Aircraft

DONG Xiaomin, YU Miao, LIAO Changrong, CHEN Weimin

2009, 22(5): 748-752.

[Abstract](1512)

Abstract:
To avoid the limitation of traditional aircraft bumpers that their stiffness and damping factors are not adjustable,a new bumper was designed.It consists of a magnetorheological (MR) damper with variable damping factors and a magnetorheological elastomer (MRE) with variable rigidity,and the two parts are connected in parallel.The control models of the two parts were derived.An adaptive variable stiffness and damping control strategy based on the energy conservation principle was proposed.Simulation results showed that the bumper made the impact energy distribute more evenly,and reduced the peak impact load by about 17%.

Structural Characteristics of New Airborne Mass Flowmeter

HUANG Danping, LIAO Junbi, LIAO Shipeng

2009, 22(5): 753-758.

[Abstract](1380)

Abstract:
The velocity vector flow of aviation fuel oil in the diver and inductive wheel of mass for airborne mass flowmeter was simulated by computational fluid dynamics (CFD) software,and the rotational moment and pressure differences of the diver and the inductive wheel of mass were calculated.Based on the calculation results,a method to increase the rotational moment and reduce the pressure difference was proposed.Then,an improved structure of the diver was designed by this method.The simulation and experiment results show that the pressure diffierence of the mass flowmeter with the improved structure was reduced by 16.0 kPa compared with that with the original structure,and met the requirements for rotational moment,pressure difference and sensitivity in airborne conditions.

Critical Traffic Volume of Traffic Signal Warrant at Major-Minor Intersections

WANG Dianhai, MA Dongfang, CHEN Yongheng

2009, 22(5): 759-763.

[Abstract](1384)

Abstract:
The feasibility of regarding multiple traffic flows as one was demonstrated by probability analysis and gap acceptance theory.On this basis,a model of average vehicle delay was established by queuing theory to find the critical traffic volume of traffic signal warrant at major-minor intersections.Then,from the viewpoint of traffic efficiency,by comparing the average vehicle delays at major-minor intersection before and after signal control,the critical volume curves of traffic signal warrant were obtained for different critical gaps of three common cross patterns at unsignalized intersection.The results may provide a theoretical basis for signal setting at major-minor intersections.

Air Traffic Flow Combinational Forecast Based on Double Gravity Model and Artificial Neural Network

ZHANG Ming, HAN Songchen, HUANG Linyuan

2009, 22(5): 764-770.

[Abstract](1851)

Abstract:
The four stage method and a double gravity model were used to forecast the OD (origin-destination) distribution of air traffic flow in the whole airspace of China.In view of the randomicity and periodicity of historical data of air traffic flow,a GM-GRNN (gray model and generalized regression neural network) combinational model was built to obtain the forecast results.The forecast results were then analyzed by Markov chain forecast model.A case study shows that,compared with the statistical data of air traffic flow OD distribution in north China airspace,the forecast results by the proposed combinational model are more precise and credible than the regression analysis and the traditional GRNN model.

Risk Identification for Driving Behaviors Based on Bayesian Discrimination

GUO Zhizheng, CHEN Chongshuang, WANG Xin

2009, 22(5): 771-775.

[Abstract](1685)

Abstract:
To effectively identify risk-state in driving behaviors,they were classified into different grades according to accident probabilities.Main factors for driving behavior were determined by single factor variance analysis.A model for risk-state identification was constructed based on Bayesian discrimination.The behavior states of five drivers were tested in a continuous driving experiment,and the measured data were divided into two groups to calibrate and test the model,respectively.The test results show that the error ratio is 4.3%.

Characteristics and Countermeasures of Traffic Accidents in Expressway Tunnel

ZHANG Yuchun, HE Chuan, WU Dexing, ZENG Yanhua

2009, 22(5): 776-781.

[Abstract](1721)

Abstract:
In older to probe into the mechanism and countermeasures of traffic accidents in expressway tunnels,10 expressway tunnels in Zhejiang Province were investigated,and the data of 599 traffic accidents and 5 280 anchorage accidents were collected.Based on these data,composition of traffic accidents,the relation between them and vehicle types and accident rate per million vehicle kilometer (MVK) were analyzed.The relationships between traffic accidents,fire accidents,rear-end accidents,wall colliding accidents and vehicle roll-over accidents with traffic volumes were investigated,and their fitting formulas were given.Some countermeasures were suggested to decrease traffic accidents in expressway tunnels.The research provides a reference for the research on traffic accident mechanism to avoid worst and catastrophic traffic accidents.

Multi-Objective Genetic Algorithm for Solving Capacitated Vehicle Routing Problems

ZOU Shurong, HUANG Xiaobin, ZHANG Hongwei

2009, 22(5): 782-786.

[Abstract](1760)

Abstract:
A multi-objective genetic algorithm based on Pareto approach was proposed for capacitated vehicle routing problems (CVRPs).In this algorithm,a new Pareto tournament selection operator based on arena’s principle is used to avoid the difficulty of solving non-convex problems;meanwhile,the nearest-neighbor algorithm and sweep algorithm are adopted to initialize population,and heuristic crossover operator is introduced to accelerate the convergence speed.The simulation result on the E-n30-k3 test specimen shows that the Pareto set obtained by this algorithm can provide manifold paths for decision makers to solve CVRPs.

New Tabu Search Algorithm for Large-Scale Vehicle Routing Problem with Simultaneous Deliveries and Pickups

LI Jian, LU Zhixiong, GAO Mourong

2009, 22(5): 787-793.

[Abstract](1466)

Abstract:
A new tabu search algorithm that integrates many neighborhood search methods,and adopts route set-based decomposition,restart and perturbation strategies,was proposed.The current solution is divided into several subsets of routes,each of which is solved by the tabu search algorithm and the best solutions of the subsets are then merged to form a new current solution.Computational results show that compared with record-to-record travel and traditional tabu search algorithm,the new tabu search algorithm can improve on 8 out of 14 of the best known solutions and reduce the number of vehicles by one in 2 cases,with error less than 2.41% for the others.

Relevance of Regional Information Industry to Economic Growth

KE Ling

2009, 22(5): 794-798.

[Abstract](1209)

Abstract:
Using multi-level interpretative structural model (ISM),analysis of the industrial relevance between information departments and other departments was made based on the 2002 input-output tables of 124 departments issued by Statistical Bureau of Sichuan Province.Then,the response coefficients and influence coefficients of 4 major industries,27 subindustries and 3 major information subindustries of Sichuan province were calculated by multiplier effect theory.The results show that the current economic growth of Sichuan province are mainly contributed by metal working and product industry,machinery manufacturing industry,construction industry,etc.,and that the leading role of the information industry have not been brought into full play yet.Therefore,it is advised that the government should make policies to vigorously develop information industry,so as to promote the economic growth of Sichuan province stably.

Search Advance Search

Journal Search

Search Type

NewsMore

instruction for authors

Read this journal

LinksMore

2009 Vol. 22, No. 5